ASSET: Analysis of Sequences of Synchronous Events in Massively Parallel Spike Trains

Abstract

With the ability to observe the activity from large numbers of neurons simultaneously using modern recording technologies, the chance to identify sub-networks involved in coordinated processing increases. Sequences of synchronous spike events (SSEs) constitute one type of such coordinated spiking that propagates activity in a temporally precise manner. The synfire chain was proposed as one potential model for such network processing. Previous work introduced a method for visualization of SSEs in massively parallel spike trains, based on an intersection matrix that contains in each entry the degree of overlap of active neurons in two corresponding time bins. Repeated SSEs are reflected in the matrix as diagonal structures of high overlap values. The method as such, however, leaves the task of identifying these diagonal structures to visual inspection rather than to a quantitative analysis. Here we present ASSET (Analysis of Sequences of Synchronous EvenTs), an improved, fully automated method which determines diagonal structures in the intersection matrix by a robust mathematical procedure. The method consists of a sequence of steps that i) assess which entries in the matrix potentially belong to a diagonal structure, ii) cluster these entries into individual diagonal structures and iii) determine the neurons composing the associated SSEs. We employ parallel point processes generated by stochastic simulations as test data to demonstrate the performance of the method under a wide range of realistic scenarios, including different types of non-stationarity of the spiking activity and different correlation structures. Finally, the ability of the method to discover SSEs is demonstrated on complex data from large network simulations with embedded synfire chains. Thus, ASSET represents an effective and efficient tool to analyze massively parallel spike data for temporal sequences of synchronous activity.

Cell assemblies at multiple time scales with arbitrary lag constellations

“…Numerous other statistical procedures for detecting assemblies or sequential patterns have been proposed previously (Grün et al, 2002a; Grün et al, 2002b; Pipa et al, 2008; Torre et al, 2016a), but most of these adhere to one or the other theoretical conceptualization of a cell assembly (cf. Figure 1A), or become computationally impractical for larger cell numbers or multiple lags (see Appendix for further discussion of both more recent and more 'traditional', cross-correlation-based, approaches).…”

“…because they rely on time-consuming bootstrap analyses (e.g. [Abeles and Gat, 2001; Pipa et al, 2008; Fujisawa et al, 2008; Gansel and Singer, 2012; Picado-Muiño et al, 2013; Torre et al, 2013, 2016a]).Along similar lines as our procedure, unitary event analysis scans simultaneously recorded spike trains for precise spike co-occurrences (Figure 1A,I) that exceed the joint spike probability predicted from independent Poisson processes with the same local rate (Grün et al, 2002a, 2002b). However, this procedure has not been extended yet to multiple lags (but see Torre et al (2016a)) or larger bins (with higher counts), and deals with non-stationarity through sliding windows or bootstrap analyses.…”

“…Numerous other statistical procedures for detecting assemblies or sequential patterns have been previously proposed (Abeles and Gerstein, 1988; Abeles and Gat, 2001, 2001b; Tetko and Villa, 2001a; Grün et al, 2002a, 2002b; Harris, 2005; Pipa et al, 2008; Sastry and Unnikrishnan, 2010; Staude et al, 2010a, 2010b; Humphries, 2011; Lopes-dos-Santos et al, 2011; Gansel and Singer, 2012; Gerstein et al, 2012; Shimazaki et al, 2012; Picado-Muiño et al, 2013; Torre et al, 2013, 2016a; Lopes-dos-Santos et al, 2013; Billeh et al, 2014; Logiaco et al, 2016), but most of these adhere to one or the other theoretical conceptualization of a cell assembly (cf. Figure 1A), or become (computationally) impractical for larger cell numbers or multiple lags, e.g.…”

Uncovering Neuronal Networks Defined by Consistent Between-Neuron Spike Timing from Neuronal Spike Recordings

“…The past decades have seen the arrival of many methods that can characterize spike timing networks (Abeles and Gerstein, 1988; Chapin and Nicolelis, 1999; Nádasdy et al, 1999; Tetko and Villa, 2001; Grün et al, 2002; Lee and Wilson, 2002; Schnitzer and Meister, 2003; Ikegaya et al, 2004; Okatan et al, 2005; Schneider et al, 2006; Nikolíc, 2007; Pipa et al, 2008; Schrader et al, 2008; Berger et al, 2010; Eldawlatly et al, 2010; Louis et al, 2010; Peyrache et al, 2010; Humphries, 2011; Lopes-dos-Santos et al, 2011; Gansel and Singer, 2012; Torre et al, 2016). Their application has led to important insights, yet they have several limitations, especially when it comes to their application on large scale neuronal recordings (Buzsáki, 2004).…”

Serial Spike Time Correlations Affect Probability Distribution of Joint Spike Events

“…In other words, whether they occur more often than is expected if the individual neurons fire independently. To investigate this issue, different approaches and methods have been taken (e.g., Aertsen et al, 1989; König, 1994; Grün et al, 1999; Grün et al, 2002a,b; Pipa and Grün, 2003; Pipa et al, 2007, 2008; Staude et al, 2010; Torre et al, 2013, 2016a). To analyze ensembles of spike trains from simultaneously recorded neurons for precise spike correlations, many of these approaches model the spike train as a Poisson process with the same rate profile as the neuron under investigation (e.g., Grün et al, 2002a,b).…”